2011한국통신학회논문지Open access

Improved CABAC Method for Lossless Image Compression

Jin Heo, Yo‐Sung Ho

Open full text 1 citations

Abstract

본 논문에서는 무손실 영상 압축의 효율을 높이가 위해 새로운 문맥 기반 적응적 이진 산술 부호화(context-based adaptive binary arithmetic coding, CABAC) 방법을 제안한다. 기존의 H.264/AVC의 CABAC은 손실 (loosy) 부호화 환경에 적합하게 설계되었기 때문에, 무손실 (lossless) 부호화 환경에서 최적의 부호화 성능을 제공하지 못한다. 따라서, 무손실 화면내 부호화 환경에서 잔여 신호의 통계적 특성을 고려하여 향상된 CABAC 기반의 잔여 신호 부호화 방법을 제안한다. 실험 결과로부터 무손실 화면내 부호화 환경에서 본 논문에서 제안하는 향상된 CABAC 방법이 기존의 CABAC 방법에 비해 평균적으로 약 18.2% 정도의 비트 수를 감소시키는 것을 확인했다. In this paper, we propose a new context-based adaptive binary arithmetic coding (CABAC) method for lossless image compression. Since the conventional CABAC in H.264/AVC was originally designed for lossy coding, it does not yield adequate performance during lossless coding. Therefore, we proposed an improved CABAC method for lossless intra coding by considering the statistical characteristics of residual data in lossless intra coding. Experimental results showed that the proposed method reduced the bit rate by 18.2%, compared to the conventional CABAC for lossless intra coding.

Open-access reader

About this research paper

What this paper is about

본 논문에서는 무손실 영상 압축의 효율을 높이가 위해 새로운 문맥 기반 적응적 이진 산술 부호화(context-based adaptive binary arithmetic coding, CABAC) 방법을 제안한다. 기존의 H.264/AVC의 CABAC은 손실 (loosy) 부호화 환경에 적합하게 설계되었기 때문에, 무손실 (lossless) 부호화 환경에서 최적의 부호화 성능을 제공하지 못한다. 따라서, 무손실 화면내 부호화 환경에서 잔여 신호의 통계적 특성을 고려하여 향상된 CABAC 기반의 잔여 신호 부호화 방법을 제안한다. 실험 결과로부터 무손실 화면내 부호화 환경에서 본 논문에서 제안하는 향상된 CABAC 방법이 기존의 CABAC 방법에 비해 평균적으로 약 18.2% 정도의 비트 수를 감소시키는 것을 확인했다. In this paper, we propose a new context-based adaptive binary arithmetic coding (CABAC) method for lossless image compression. Since the conventional CABAC in H.264/AVC was originally designed for lossy coding, it does not yield adequate performance during lossless coding. Therefore, we proposed an improved CABAC method for lossless intra coding by considering the statistical characteristics of residual data in lossless intra coding. Experimental results showed that the proposed method reduced the bit rate by 18.2%, compared to the conventional CABAC for lossless intra coding.

Why it matters

OpenAlex reports 1 citations for this work. Citation counts describe recorded attention and do not establish research quality.

Key contribution

A contribution statement is not available in the OpenAlex record.

Method / approach

Method details are not available in the OpenAlex metadata.

Main findings

Findings are not separately available in the OpenAlex metadata.

Limitations

Limitations are not available in the OpenAlex metadata.

Applications

Application details are not available in the OpenAlex metadata.

Available abstract

본 논문에서는 무손실 영상 압축의 효율을 높이가 위해 새로운 문맥 기반 적응적 이진 산술 부호화(context-based adaptive binary arithmetic coding, CABAC) 방법을 제안한다. 기존의 H.264/AVC의 CABAC은 손실 (loosy) 부호화 환경에 적합하게 설계되었기 때문에, 무손실 (lossless) 부호화 환경에서 최적의 부호화 성능을 제공하지 못한다. 따라서, 무손실 화면내 부호화 환경에서 잔여 신호의 통계적 특성을 고려하여 향상된 CABAC 기반의 잔여 신호 부호화 방법을 제안한다. 실험 결과로부터 무손실 화면내 부호화 환경에서 본 논문에서 제안하는 향상된 CABAC 방법이 기존의 CABAC 방법에 비해 평균적으로 약 18.2% 정도의 비트 수를 감소시키는 것을 확인했다. In this paper, we propose a new context-based adaptive binary arithmetic coding (CABAC) method for lossless image compression. Since the conventional CABAC in H.264/AVC was originally designed for lossy coding, it does not yield adequate performance during lossless coding. Therefore, we proposed an improved CABAC method for lossless intra coding by considering the statistical characteristics of residual data in lossless intra coding. Experimental results showed that the proposed method reduced the bit rate by 18.2%, compared to the conventional CABAC for lossless intra coding.

Key concepts: Context-adaptive binary arithmetic coding, Context-adaptive variable-length coding, Lossless compression, Lossy compression, Adaptive coding, Arithmetic coding, Tunstall coding, Computer science

Related papers

Back to paper searchBrowse research topicsOriginal source
Improved CABAC Method for Lossless Image Compression — Research Paper | ScholarLens